In the Wiktionary projects the pronunciation is an aspect that is often given attention. Pronunciation are something that we would like to have with every word and phrase.
How to represent pronunciations
In the lexicological world the traditional way of working with pronunciation is either by transcription to the character set of the language or by transcribing it in a phonetic script. There are inherent problems with all these options. Transcriptions are often done once and not for the language that the user speaks. The phonetic scripts are not universally understood and, the way they are implemented is often language dependent.
The best way to represent pronunciations is by offering them as a soundfile.
How to save soundfiles for usage in the Wikimedia projects
As every Wiktionary project aims to describe every word of every language, the pronunciation soundfiles are saved in Wikimedia Commons. It is vitally important that the content is saved with appropriate license information. Without it, this content will be removed from Commons. When a pronunciation file is uploaded, it is expected to conform to a naming convention.
Wikimédia France developed a project, Lingua Libre , which allows to easily record many words of any language in a short amount of time (several hundred words per hour). The files are automatically uploaded to Wikimedia Commons with metadata about the speaker (like their place of residence and their level of language proficiency).
Basic audio recording workflow
Audio recording workflow with Lingua Libre
Local wiktionary standards
As many wiktionaries do not cooperate there are many different ways of including the pronunciations on the wiktionary itself. Please have a look how this is done on a particular wiktionary.